This study is a supplement to an existing R01 study of the life course and genetic structure of autistic traits (HD-042541). Although autism is largely an inherited disorder, the specific genes that cause most cases of the disorder remain unknown. Previous genetic studies of autism, which primarily involve affected pairs of siblings, have yielded only modest associations with specific chromosomal regions. The affected sib pair design, however, has important limitations. Affected sib pairs are relatively rare, and this limits the sample size that can be used to search for genes for autism. For complex diseases like autism, in which numerous genes may be acting in concert to cause the disorder, larger sample sizes may be needed to identify the genes (or combinations of genes) that are responsible. One way to increase both the size of a sample and its power to detect genes, is to adopt a quantitative genetic approach, in which genetic markers are related not just to presence or absence of the disorder, but to the degree of symptom severity that each individual manifests. Using information on all family members, including those with only subtle autistic-like traits, we have shown in an initial pilot sample of 125 families that the ability to identify genes for autism using a quantitative approach may be more powerful than traditional affected sib pair methods. We are currently in the process of enlarging the sample to a total of 255 families, in which all siblings in each family are being quantitatively phenotyped. This is a proposal to further enhance the statistical power of this quantiative linkage effort by 1) obtaining quantitative assessments of autistic social impairment (the "phenotype" of interest) on the parents in each family; and 2) enhancing the density of genotypic markers for the entire sample of children and their parents (n=255 families). If the original findings are replicated and the linkage regions refined, they would directly narrow the search for genes in several chromosomal regions which are currently suspected of harboring autism susceptibility genes. Furthermore, these quantitative methods could be applied to large collections of extended autism pedigrees and could greatly enhance the power to identify other genes of major effect in autism. [unreadable] [unreadable] [unreadable] [unreadable]